Missing data is a common problem in longitudinal data. Methods for handling incomplete longitudinal data often depend on the pattern of missingness and the mechanism that generates the missing values. The missing data pattern is intermittent whenever an observed value is available even after a missing value occurs. A mechanism where missingness depends on the unobserved data perhaps in addition to the observed data, is non- random. Ignoring the missing values with such data would lead to biased conclusion. Estimating parameters with nonrandom missing data is complex. Likelihood based methods require specification of the joint distribution of the data and the missing data mechanism. This specification can be further classified into pattern-mixture and selection models. There exists a further classification of the models into marginal model and random-effects model. In this paper, normal random effects model is used to fit longitudinal data in the presence of intermittent nonrandom missing values. The stochastic EM algorithm is developed to obtain the model parameter estimates Also, missingness parameters have been obtained. The bootstrap method has been used to compute the standard errors of estimates. The proposed method is applied to a data set concerning quality of life among breast cancer patients in a clinical trial undertaken by the International Breast Cancer Study Group.